Generative AI Image Models Perpetuate Racial Stereotypes in Worker Depictions

Thumbnail Image

The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.

Generative AI systems like Stable Diffusion and DALL-E have been found to produce biased images when prompted for 'African workers' versus 'European workers.' The outputs reinforce harmful stereotypes, depicting Africans in poverty and Europeans as affluent, highlighting the real-world harm caused by AI-driven bias and discrimination.[AI generated]

Why's our monitor labelling this an incident or hazard?

The article explicitly discusses AI systems (Stable Diffusion, DALL-E) generating biased images that reinforce harmful stereotypes about African workers compared to European workers. This bias is a direct consequence of the AI systems' training and use, leading to harm in the form of perpetuating racial stereotypes and discrimination, which falls under violations of human rights and harm to communities. The harm is realized and ongoing as the biased images are generated and disseminated. Hence, this qualifies as an AI Incident.[AI generated]
AI principles
FairnessRespect of human rightsTransparency & explainabilityAccountabilityHuman wellbeingSafety

Industries
Media, social platforms, and marketingArts, entertainment, and recreation

Affected stakeholders
General public

Harm types
Human or fundamental rightsReputationalPsychological

Business function:
Research and developmentMarketing and advertisement

AI system task:
Content generation


Articles about this incident or hazard